Research On De-multiplexing Algorithms For PDM And MGDM Systems

Polarization division multiplexing(PDM)technique aims at creating two independent communication channels over the transmission wavelength in the single mode fiber by using the independent and mutually orthogonal states of polarization,and the mode group diversity multiplexing(MGDM)technique can use different mode groups in the multimode fiber as different and independent transmission channels.Both PDM and MGDM technologies can not only increase the capacity of optical fiber communication system and spectrum efficiency,but also upgrade system rapidly and economically on the existing optical network.While the two polarizations in the PDM system can be affected by polarization mode dispersion,polarization dependent loss etc.,which will cause crosstalk between the two polarizations and similarly in the MGDM system modal dispersion in a multimode fiber will lead to crosstalk among the channels.The crosstalk makes the received signals are weighted superposition of source signals and degrades the transmission performance of communication systems.Therefore,eliminating the crosstalk existing in the received signals to realize de-multiplexing with digital signal processing methods are a hot area of research for the two systems.This paper focuses on studying the polarization de-multiplexing algorithms in PDM system and preliminarily discusses the de-multiplexing technology in MGDM system at the same time.The basis of the major algorithms used here-independent component analysis(ICA)algorithm is introduced in detail in this paper.After studying the FastICA algorithm only used for real-valued signals separation,it demonstrates the effectiveness of separating real-valued signals for FastICA by observing waveforms in the time domain simulated by MATLAB.And by analyzing the characteristics of complex signals received by polarization diversity coherent receiver after transmitted through the single mode fiber the twice real-valued FastICA(TR-FastICA)algorithm is proposed in this paper to de-multiplex complex observed signals in CO-PDM system.The different computational complexity in the proposed algorithm and one kind of complex ICA algorithms-complex maximization of non-Gaussianity(CMN)algorithm is compared and analyzed through theoretical derivation and it comes to the conclusion that the proposed algorithm can reduce computationalcomplexity and can de-multiplex complex signals effectively through analyzing the results with and without noise by MATLAB.The constellations and bit error rate(BER)curve graphs cosimulated by MATLAB and VPI further demonstrate the effectiveness of de-multiplexing.Chaotic particle swarm optimization(CPSO)algorithm is combined with ICA to realize de-multiplexing.The constellations of signals with and without noise obtained by MATLAB and distribution situations of particles and iteration curve graphs cosimulated by MATLAB and VPI prove that CPSO-ICA can de-multiplex complex signals with good convergence performance.The polarization de-multiplexing performance of CPSO-ICA and TR-FastICA are compared by observing BER results and the conclusion is that TR-FastICA performs better than CPSO-ICA for x polarization,however CPSO-ICA performs better than TR-FastICA for y polarization.MGDM technology is also introduced in this paper and FastICA is applied in a 3×3NRZ-OOK-MGDM system to perform mode group de-multiplexing and it comes to the conclusion through analyzing results simulated by MATLAB that FastICA can de-multiplex three multiplexed signals with crosstalk at some degree.